Data can provide enlightening information to a company’s leaders, guiding their decisions. But measurements can be misleading, which will push you in the wrong direction. Identifying whether your data can be trusted isn’t as difficult as you may think. Start by clarifying what you want to know. If you are timing a multistep process, for example, do you want to know how long each individual step is or how long the entire process is? Both have valid uses but are quite different. Then see how closely the measurements align with what you really want to know. Strive to distinguish “pretty close” from “a good-enough indicator” from “not what I had in mind.” You may have to settle for a less-than-perfect measurement, but at least you’ll know exactly what it is you are settling on. And be sure to try the “smell test.” When results just don’t seem right, dig deeper.
Adapted from “4 Steps for Thinking Critically About Data Measurements,” by Thomas C. Redman